Abstract:Dictionary-based classification has been promising in knowledge discovery from image data, due to its good performance and interpretable theoretical system. Dictionary learning effectively supports both small- and large-scale datasets, while its robustness and performance depends on the atoms of the dictionary most of the time. Empirically, using a large number of atoms is helpful to obtain a robust classification, while robustness cannot be ensured when setting a small number of atoms. However, learning a huge dictionary dramatically slows down the speed of classification, which is especially worse on the large-scale datasets. To address the problem, we propose a Fast and Robust Dictionary-based Classification (FRDC) framework, which fully utilizes the learned dictionary for classification by staging - and -norms to obtain a robust sparse representation. The new objective function, on the one hand, introduces an additional -norm term upon the conventional -norm optimization, which generates a more robust classification. On the other hand, the optimization based on both - and -norms is solved in two stages, which is much easier and faster than current solutions. In this way, even when using a limited size of dictionary, which makes sure the classification runs very fast, it still can gain higher robustness for multiple types of image data. The optimization is then theoretically analyzed in a new formulation, close but distinct to elastic-net, to prove it is crucial to improve the performance under the premise of robustness. According to our extensive experiments conducted on four image datasets for face and object classification, FRDC keeps generating a robust classification no matter whether using a small or large number of atoms. This guarantees a fast and robust dictionary-based image classification. Furthermore, when simply using deep features extracted via some popular pre-trained neural networks, it outperforms many state-of-the-art methods on the specific datasets.

Scale Adaptive Dictionary Learning

Online Discriminative Dictionary Learning for Image Classification Based on Block-Coordinate Descent Method

Multiscale Dictionary Learning for Hierarchical Sparse Representation.

Learning a Collaborative Multiscale Dictionary Based on Robust Empirical Mode Decomposition

Adaptive Submodular Dictionary Selection for Sparse Representation Modeling with Application to Image Super-Resolution

Double Sparse Dictionary Learning for Image Super Resolution

Low Rank Multi-Dictionary Selection at Scale

An augmented Lagrangian multi-scale dictionary learning algorithm

Large-scale Dictionary Learning for Local Coordinate Coding.

Multi-level Discriminative Dictionary Learning with Application to Large Scale Image Classification.

Multiscale Online Dictionary Learning for Quality Scalable Video Coding

Multiscale Adaptive Representation Of Signals: I. The Basic Framework

Multi-scale Dictionary for Single Image Super-Resolution

Self-paced Dictionary Learning for Image Classification

Multiscale Incremental Dictionary Learning with Label Constraint for SAR Object Recognition.

Atom dimension adaptation for infinite set dictionary learning

Discriminative Dictionary Pair Learning With Scale-Constrained Structured Representation for Image Classification

Efficient Dictionary Learning with Sparseness-Enforcing Projections

A Dictionary Approach to Domain-Invariant Learning in Deep Networks

Active Dictionary Learning in Sparse Representation Based Classification

Fast and Robust Dictionary-based Classification for Image Data